Overview
TextIn OCR MCP Server is a modular MCP server that enables OCR-powered text extraction from documents, including document text recognition, ID recognition, and invoice recognition, with support for converting documents into Markdown format. It exposes three tools: recognition_text, which performs text recognition on images, Word documents, and PDFs; doc_to_markdown, which converts PDFs, Word documents, and common image formats to Markdown; and general_information_extration, which automatically identifies and extracts key information from documents or user-specified fields, returning a JSON with the extracted data. When inputs are provided via URL, access to protected resources is not supported. The project provides a setup path based on NPX usage, requiring an APP_ID and APP_SECRET from TextIn, and a MCP server configuration that includes environment variables like APP_ID, APP_SECRET, and MCP_SERVER_REQUEST_TIMEOUT. Supported input formats include PDF, Word/Excel documents, and standard image types (JPEG, PNG, BMP). The MCP server is released under the MIT License, enabling usage, modification, and distribution subject to license terms.
Features
recognition_text
Text recognition from images, Word documents, and PDFs; returns extracted text.
doc_to_markdown
Converts images, PDFs, and Word documents to Markdown; returns Markdown text.
general_information_extration
Automatically identifies and extracts key information from documents or user-specified fields; returns a JSON with the extracted data; supports PDF, Word/Excel, and images.
Who Is This For?
- Developers:Integrate OCR, text extraction, and Markdown conversion into applications via the MCP server.
- Data engineers:Extract structured data from documents and generate Markdown representations within pipelines.
